GPT-5 was connected to Ginkgo Bioworks' automated cloud laboratory to optimize cell-free protein synthesis through closed-loop experimentation. Over six rounds testing 36,000+ reactions, the system achieved a 40% reduction in protein production costs compared to previous best results. The autonomous lab setup allowed GPT-5 to propose experiments, analyze results, and iteratively refine reaction compositions, discovering novel combinations that work well under high-throughput automation constraints. The approach demonstrates how AI models can accelerate biological research by removing iteration bottlenecks when directly integrated with lab automation.
Table of contents
Why cell-free protein synthesis mattersCell-free protein synthesis is difficult to optimize and costlyConnecting GPT‑5 to a robotic labWhat we learnedLimitationsWhat’s nextSort: